The data that follows was collected and analyzed by Victor Amin of SendGrid’s Big Data team and summarized by Jacob Hansen from the Delivery team.
When marketers are deciding what criteria they should use to measure the health or potential deliverability success of their email address list, we know that it is hard to narrow down that decision to just one or two indicators. Even a handful of known indicators would be nice. But, we see that there are numerous factors to consider at all times.
One main factor we have seen recently as a great indicator of list health is the presence of “role addresses.” They may indicate poor list acquisition practices or poor list management practices. Either way, one of our SendGrid data scientists, Victor Amin, was able to prove that they do have a negative impact on email delivery. His findings are shared below.
What Role Addresses Are
Role addresses are email addresses that typically define a responsibility rather than a person. Therefore, they are unlikely to give their consent to receive marketing emails. Role addresses can fall into a few subcategories:
- “Broadcast” addresses: Emails to these addresses go to many people at once.(examples: support@ or all@)
- “Position” addresses: These addresses might go to an individual, but the individual is not the owner of the address and the human recipient can likely change. (examples: admin@ or ceo@)
- “Service” addresses: Service addresses are used as an interface to a program or service. Unwanted emails to them can actually incur a significant cost to the operator. For instance, this could be a service address that faxes anything emailed to it long distance. (examples: ftp@ and uucp@)
- “Test” addresses: These addresses are generally not published and are reserved for internal testing. (examples: test@ or asdf@)
- “Compliance” addresses: Compliance addresses are designed to identify bad senders. (examples: spam@ or abuse@)
Impact On Delivery and Reputation
Having role addresses on newsletter lists is considered a bad practice in general. But, there hasn’t been much data available to back up that claim. We recently developed an empirical method to find role addresses and review their effect on deliverability. We quantified the effects of role addresses on bounces, unsubscribes, spam reports, delivery, and engagement. What we found bolsters the claim that role addresses on a list have a negative effect: even a single role address on a newsletter list is associated with more bounces, more unsubscribes, and decreased delivery and engagement rates.
The plot below shows the median delivery rate (blue line), along with the 75th and 25th percentile delivery rates (light blue area) with respect to the number of role addresses. As suspected, a higher count of role addresses is correlated with lower delivery rates. And, while it may seem that the drop from 100% to 98% isn’t that drastic, remember that this is data pulled from 580,087 newsletters sent through SendGrid and 514,920,776 individual emails.
This additional graph below shows the results from when we reviewed the presence of role addresses within all email messages compared to their larger presence in “Bounce” and “Unsubscribe” groups. We found that role addresses account for about 0.49% of all newsletter send requests. However, they account for 1.3% of unsubscribes and 1.6% of bounces (2.7x and 3.3x greater than the baseline send rate).
Creating Your Role Address List to Check Against
It’s often unclear what counts as a role address—different sources have different lists. And, no two role address lists are the same. Within SendGrid’s research method, we found 91 previously unreported role addresses. These were mostly foreign language role addresses. When combined with SendGrid’s current list, that gives us a comprehensive list of 251 role addresses. We recommend validating your own list against our new comprehensive list of role addresses listed below.
Remember that the presence of these role addresses are most likely an indicator of a larger list acquisition or list hygiene issue that is worth exploring deeper. But, it is definitely a good starting point.
SendGrid’s Full Role Address List:
For more from Victor and the Big Data team, check out their recent study on email engagement and subject lines and for more delivery tips from Jacob, read his latest blog posts.